Problems

This site is a place to put thoughts.

I think about problems first, before solutions, before products, before companies. A problem is a constraint, and constraints are what give a solution shape. Creating something without a real problem to point at feels futile to me.

Most of what shows up here is in that mode. I pick a question that nobody seems to have agreed is the question yet — usually because it's getting answered by default, in code, before anyone names it — and try to work out what's actually being decided. The questions I care about most right now are about the agent transition: who owns the data, what the data substrate has to look like, what a different default could ship as.

I also think about physics. Simulation theory, cosmology, the structure of physical law. It looks like a tangent, but the same problem-first thinking shows up there, and pretending otherwise got boring fast. Some pieces here will be physics. Most won't.

I bought problems.vc a long time ago. The idea was to build an agentic research organization that scoured the internet for the biggest, hardest, most important problems still standing — somewhere a person could go to find a problem worth working on. I haven't built that yet. But the underlying belief is the same one I still operate from: helping each other find the right problems is itself the work. Most things fail not because the execution was bad but because the problem was wrong, or never named clearly, or ten thousand people were already chasing it.

So this is a notebook. It's where I think out loud about problems worth thinking about, and occasionally about what I'm building next to one of them. Pieces ship when they're worth publishing.

Subscribe via RSS or email if any of this resonates.

— Matt